1 of 34

Modeling viral dynamics of SARS-CoV2:�treatment & transmission

Jérémie Guedj

INSERM UMR 1137, Paris, France

jeremie.guedj@inserm.fr

2 of 34

Viral dynamic modeling to support 3 main projects

January

February

March

March 22^nd

February 5^th

1st patient included in the French Covid

prospective cohort (PI: Jade Ghosn)

Phase III controlled trial of the safety and efficacy of treatments of COVID-19 in hospitalized adults (PI: Florence Ader)

Early March

Development of a NHP model of SARS-CoV-2 infection (PI: Roger Le Grand)

Can we use viral dynamics to better understand the determinants of virus transmission and to optimize the use of antiviral therapy ?

As everyone here knows, an unprecedented outbreak of Ebola infection has been ravaging Guinea, Liberia and Sierra Leone since December 2013.

In Guinea, more than 3000 cases have occurred so far, of whom 2000 died.

The disease has very high mortality rates in West Africa, even in experienced ebola care centers that use a comprehensive package of supportive care .

There is thus an urgent need to find new curative interventions.

Several agents have been identified as having a specific activity against Ebola. Favipiravir, a pyrazine carboximide derivative, is one of them. This drug has proven effective and well tolerated in humans with influenza. It has also been shown to inhibit EBOLA virus replication in vitro, and to decrease mortality in a mouse model of EBOLA virus infection.

Favipiravir was one of the promising agents identified by the WHO for fast-track testing

3 of 34

Viral dynamics during acute infection

Peak viral load ≈ Target cell exhaustion

Exponential growth of infected cells

Elimination of infected cells

Bonhoeffer et al, PNAS (1996)

Credit: NIAID

As everyone here knows, an unprecedented outbreak of Ebola infection has been ravaging Guinea, Liberia and Sierra Leone since December 2013.

In Guinea, more than 3000 cases have occurred so far, of whom 2000 died.

The disease has very high mortality rates in West Africa, even in experienced ebola care centers that use a comprehensive package of supportive care .

There is thus an urgent need to find new curative interventions.

Several agents have been identified as having a specific activity against Ebola. Favipiravir, a pyrazine carboximide derivative, is one of them. This drug has proven effective and well tolerated in humans with influenza. It has also been shown to inhibit EBOLA virus replication in vitro, and to decrease mortality in a mouse model of EBOLA virus infection.

Favipiravir was one of the promising agents identified by the WHO for fast-track testing

4 of 34

Timing of antiviral treatment is key to avoid disease progression

As everyone here knows, an unprecedented outbreak of Ebola infection has been ravaging Guinea, Liberia and Sierra Leone since December 2013.

In Guinea, more than 3000 cases have occurred so far, of whom 2000 died.

The disease has very high mortality rates in West Africa, even in experienced ebola care centers that use a comprehensive package of supportive care .

There is thus an urgent need to find new curative interventions.

Several agents have been identified as having a specific activity against Ebola. Favipiravir, a pyrazine carboximide derivative, is one of them. This drug has proven effective and well tolerated in humans with influenza. It has also been shown to inhibit EBOLA virus replication in vitro, and to decrease mortality in a mouse model of EBOLA virus infection.

Favipiravir was one of the promising agents identified by the WHO for fast-track testing

5 of 34

SARS-CoV-2 viral dynamics in mild patients

Young et al, JAMA 2020 ; Gonçalves et al, CPT:PSP (2020)

Treatment needs to be administarted early and blocks more than 90% of viral infection/production to have a dramatic effects on the course of the disease

- prevent infection (if administrated as P(R)EP)

- prevent symptomatic infection

As everyone here knows, an unprecedented outbreak of Ebola infection has been ravaging Guinea, Liberia and Sierra Leone since December 2013.

In Guinea, more than 3000 cases have occurred so far, of whom 2000 died.

The disease has very high mortality rates in West Africa, even in experienced ebola care centers that use a comprehensive package of supportive care .

There is thus an urgent need to find new curative interventions.

Several agents have been identified as having a specific activity against Ebola. Favipiravir, a pyrazine carboximide derivative, is one of them. This drug has proven effective and well tolerated in humans with influenza. It has also been shown to inhibit EBOLA virus replication in vitro, and to decrease mortality in a mouse model of EBOLA virus infection.

Favipiravir was one of the promising agents identified by the WHO for fast-track testing

6 of 34

Peak of viral infection coincides with symptom onset (prior to Omicron and vaccine)

Néant et al, PNAS (2021)

As everyone here knows, an unprecedented outbreak of Ebola infection has been ravaging Guinea, Liberia and Sierra Leone since December 2013.

In Guinea, more than 3000 cases have occurred so far, of whom 2000 died.

The disease has very high mortality rates in West Africa, even in experienced ebola care centers that use a comprehensive package of supportive care .

There is thus an urgent need to find new curative interventions.

Several agents have been identified as having a specific activity against Ebola. Favipiravir, a pyrazine carboximide derivative, is one of them. This drug has proven effective and well tolerated in humans with influenza. It has also been shown to inhibit EBOLA virus replication in vitro, and to decrease mortality in a mouse model of EBOLA virus infection.

Favipiravir was one of the promising agents identified by the WHO for fast-track testing

7 of 34

Viral dynamics and transmission

Viral dynamics and early treatment

Viral dynamics in hospitalized patients

Outline

8 of 34

Viral load as an « usual suspect » of infectiousness

Van Kampen et al, Nature Comm 2021; Néant et al, PNAS 2021 ; Jones et al, Science 2021

9 of 34

Variants of concern are associated with a higher viral load

Jones et al, Science (2021); Naveca et al, Nature Medicine (2021) ; Chia et al, submitted

Blanquart et al, Journal of Infection (2021); Blanquart et al, submitted ; Ke et al, MedrXiv (2021)

10 of 34

Vaccination reduces viral load by 4-50 fold in infected individuals (2-6 Ct) (at least prior to delta variant)

Levine-Tiefenbrun et al, Nature Medicine (2021) ; Thompson et al, New England Journal of Medicine (2021)

11 of 34

Diagnosis D0

D3

D7

High risk contact

Viral load is associated with the risk of transmission

Marks et al, Lancet Infectious Diseases (2021)

12 of 34

259 index cases with available viral load data and 582 high risk contacts were retained in our study

The first swab was obtained at a median time of 4 days post symptom onset

60% of contacts were households and 40% were non-households

The rate of transmission in household was 24.9% (87/349)
non-household was 12.4% (29/233)

Brief data description

13 of 34

Most contacts occured arround symptoms onset

High-risk contact that has led to an infection

High-risk contact that has NOT led to an infection

14 of 34

Joint modelling of viral load and probability of transmission

15 of 34

15

Joint modelling of viral load and probability of transmission

16 of 34

16

Threshold for viral culture [3]

[3] Néant et al, PNAS, 2021

Joint modelling of viral load and probability of transmission

17 of 34

17

Threshold for viral culture [3]

[3] Néant et al, PNAS, 2021

Joint modelling of viral load and probability of transmission

18 of 34

18

Strength of the association between viral load and probability of transmission

Threshold for viral culture [3]

[3] Néant et al, PNAS, 2021

Joint modelling of viral load and probability of transmission

19 of 34

19

Strength of the association between viral load and probability of transmission

Viral load at the time of contact

Threshold for viral culture [3]

[3] Néant et al, PNAS, 2021

Joint modelling of viral load and probability of transmission

20 of 34

20

Several Models tested:

M1: No effect of viral load on transmission

M2: Logit-linear effect of viral load

M3: Log-linear effect of viral load

SAEM algorithm

Monolix 2020

[3] Néant et al, PNAS, 2021

Joint modelling of viral load and probability of transmission

21 of 34

21

Reconstructed viral dynamics in index cases

High-risk contact that has led to an infection

High-risk contact that has NOT led to an infection

Parameter estimation using a lognormal distribution of the incubation period with a mean value of 5 days

22 of 34

Modelling the probability of transmission

Parameter estimates (RSE %)
No effect of viral load (M1)		Logit-Linear (M2)		Log-Linear Model (M3)
Fixed effect	Random effect SD	Fixed effect	Random effect SD	Fixed effect	Random effect SD
5	0.125	5	0.125	5	0.125
				13.40 (22)	0.423 (35)
0.83 (1)				0.832 (100)	0.023 (74)
					2.3 (9)
1.28 (38)		0.49 (20)		0.47 (6)	0.545 (23)
0.57 (62)		0.21 (44)		0.25 (17)	0.545 (23)
2502		2497		2500

23 of 34

Modeling the probability of transmission over time

Model based predictions of the dynamics of viral load and infectiousness over time. Prediction interval of the viral load (black) and the probability of transmission over time after a high-risk contact obtained from 1,000 simulations of the model. The shaded area represents the 90% inter quantile range.

Risk of infection after a high-risk contact in household

Risk of infection after a high-risk contact not in household

Peak transmission probability (90% IQR):

Household : 29% [6-96]
Non-household : 13% [5-38]

24 of 34

Modeling the probability of transmission over time

25 of 34

25

Sampling the generation time

Time elapsed between the infection of an individual and the infection of a contact

Sampled an simulated individual and a time of contact
Contact outcome (infection or not) based of the transmission probability at the time of contact
500 000 iterations

Cerada et al, submitted, 2020

Bi et al, The Lancet. Infectious Disease, 2020

26 of 34

26

Time elapsed between the infection of an individual and the infection of a contact

We considered 2 types of distribution of contacts

Time-varying distribution

Cerada et al, submitted, 2020

Bi et al, The Lancet. Infectious Disease, 2020

Density

Sampling the generation time

27 of 34

27

Time elapsed between the infection of an individual and the infection of a contact

We considered 2 types of distribution of contacts

Time-varying distribution
Constant distribution

Cerada et al, submitted, 2020

Bi et al, The Lancet. Infectious Disease, 2020

Sampling the generation time

28 of 34

28

Time elapsed between the infection of an individual and the infection of a contact

We considered 2 types of distribution of contacts

Time-varying distribution
Constant distribution

Time-varying – median [90% IQR]:

Household: 5.1 days [1-10]
Non-household: 4.8 days [1-11]

Constant:

Household: 7.7 days [2-17]
Non-household: 8.2 days [2-18]

Cereda et al, submitted, 2020

Bi et al, The Lancet. Infectious Disease, 2020

Sampling the generation time

29 of 34

Impacts of variants or vaccination on viral load are key for transmission

30 of 34

30

The effect of variants on viral load

4 fold increase of viral production

Household: 42% increase in transmission
Non-household: 27% increase in transmission

8 fold reduction of viral production

Household: 50% reduction in transmission
Non-household: 75% reduction in transmission

31 of 34

Kissler et al, NEJM (2021) ; Ke et al., MedrXiv (2021) ; Despres et al., MedrXiv (2021) ; Eyre et al., NEJM (2022)

Viral load : infectiousness may be variant and vaccine dependent

Vaccine may not be as effective against Delta than against other variants

New data constantly emerge: still a story to build !

32 of 34

33 of 34

Acknowledgements

French Covid & Discovery study groups

Florence Ader, INSERM

Drifa Belhadi, INSERM

Maude Bouscambert, INSERM

Charles Burdet, INSERM

Helene Espérou, INSERM

Jade Ghosn, INSERM

Cédric Laouénan, INSERM

Quentin Le Hingrat, INSERM

Yazdan Yazdanpanah, INSERM

The Reacting pre-clinical research group

Vanessa Contreras, CEA

Xavier de Lamballerie, AP-HM

Roger Le Grand, CEA

Pauline Maisonnasse, CEA

Romain Marlin, CEA

Manuel Rosa Calatrava, INSERM

Caroline Solas, AP-HM

Sylvie Van der Werf, Pasteur

Modeling

François Blanquart, INSERM

Alan S. Perelson, Los Alamos National Laboratory

Steven Kern, BMGF

Ping Zhao, BMGF

Colin Pillai, CP+

Patrick Smith, Certara

Michael Marks, London School of Tropical Medicine

PhD candidates and Postdocs

Pete Czuppon, College de France

Antonio Gonçalves, INSERM (now Certara)

Guillaume Lingas, INSERM

Aurélien Marc, INSERM

Nadège Néant, INSERM

34 of 34

Join our group for a Postdoc in Paris !

For more information: jeremie.guedj@inserm.fr